Current social and demographic trends, combined with 'the new policy agenda', highlight the importance of nurses' role in educating patients about medication. In the absence of previous research investigation, this study set out to explore nurses' current contribution to medication education and the clinical contextual factors that influence current practice. The evidence base for effective medication education was established from reviews of literature and focus groups with key informants. Nurses' practice was investigated using a case study approach in seven clinical areas representing adult, care of the older person, mental health and community nursing contexts. Methods used to collect data were: audio-recordings (n=37) and observation (n=48) of nurse-patient interactions about medication, post-interaction interviews with nurses (n=29), post-interaction interviews with patients (n=39), analysis of relevant written documentation and researcher observation and field notes. Data sources within each case were subjected to systematic content analysis in order to identify current practice and contextual influences within each case. Cross-case analysis was also employed in order to identify explanations for any differentiation in practice. Findings indicate that nurses' contribution to medication education is commonly limited to simple information giving about medicines, involving the name, purpose, colour, number of tablets and the time and frequency that medications should be administered. Nurses' practice in two of the seven clinical areas was characterised by interactions that more closely demonstrated features of what is known to constitute more comprehensive and effective medication education. Analysis of contextual influences within and between cases allowed explanations to be derived for the types of medication education interactions observed. These concerned: patient characteristics, perceived and expressed preferences of patients for information, characteristics of the nurse-patient relationship, lack of time and high workload, and the philosophy of care within the clinical area. In all clinical areas, nurses were not explicitly and judiciously using available evidence to inform their medication-related interactions. The paper concludes with discussion and implications of the findings.